ISSN: 2038-7296POLIS Working Papers

[Online]

Istituto di Politiche Pubbliche e Scelte Collettive – POLISInstitute of Public Policy and Public Choice – POLIS

POLIS Working Papers n. 205

October 2013

The choice to enrol in a small university:A case study of Piemonte Orientale

Tiziana Caliman and Alberto Cassone

UNIVERSITA’ DEL PIEMONTE ORIENTALE “Amedeo Avogadro” ALESSANDRIA

Periodico mensile on-line "POLIS Working Papers" - Iscrizione n.591 del 12/05/2006 - Tribunale di Alessandria

1

The Choice to Enrol in a Small University:

A Case Study of Piemonte Orientale

Tiziana Caliman* and Alberto Cassone **

Abstract1

In the recent past, expectations concerning universities have emphasised their active role in

enhancing economic and regional development. The universities in geographical areas suffering from

structural problems are particularly required to play this role. Moreover, the correlation between the

socioeconomic status (and the education) of parents and that of their adult offspring is positive and

significant, in both the statistical and practical senses. This paper investigates the experience of a

small Italian University (Piemonte Orientale “Amedeo Avogadro”), in order to evaluate its role in

human capital accumulation, necessary to economic development. The aim of this article is to verify

whether this small university satisfies a specific demand which would never be satisfied by a larger

university. We found important role of small Universities in the human capital accumulation in the

recruitment basin, a phenomenon with medium and long term implications. The empirical results show

that the representative graduate student of Piemonte Orientale is characterized by modest parental

socioeconomic conditions and education. Its demographic recruitment basin is a specific and well

defined geographical area. These factors have a positive impact on the choice of enrolment (Piemonte

Orientale versus other Universities). The choice is modelled by a probit (logit) binary outcomes model

using the Almalaurea cross-section sample on graduates in year 2008. We also update the dataset and

re-estimate the models in order to verify the robustness of empirical results and to identify changes in

the representative student, using the Almalaurea cross-section sample on graduates for year 2010.

The crucial role of the modest socioeconomic background and the low mobility of the students are

confirmed. The 2010 analysis does not confirm a result for year 2008, i.e. that a poor performance in

secondary school increases the probability to choose Piemonte Orientale vs larger and well

1 The paper has been conceived and written together, while Tiziana Caliman is responsible for the econometric analysis.

This paper is the result among others, of a research project supported by the University of Piemonte Orientale A. Avogadro –

CESPA, including a grant sufficient for the acquisition data from AlmaLaurea. The authors express their gratitude to

prof.Andrea Cammelli (Alma Laurea Director) and particularly to dr.Angelo di Francia (Alma Laurea scientific staff) for

assistance and comments.

The authors are also indebted to colleagues for comments during the presentation at POLIS Dept. Seminar ,Spring 2011

and Winter 2013.

2

established universities: the result underlines a positive evolution of this small university recruitment

performance.

Keywords: Performance; Human Capital Accumulation; Small Universities

JEL Classification: I20, I21, I23, R00

* CerTet, Bocconi

Via Roentgen 1,20136 Milan, Italy

e-mail: caliman@economia.unige.it

* * Institute of Public Policy and Public Choice – POLIS, University of Piemonte Orientale

Via Cavour 84, 15121 Alessandria, Italy

e-mail: alberto.cassone@sp.unipmn.it

1. Introduction

As in many EU member countries, and in the particular context of the Bologna Convention on

Tertiary Education, the Italian university system has experienced substantial reforms in recent years2.

The key reform aims include increasing the participation, progression and retention rates of students in

higher education. Reform has reduced, on the one hand, the length of undergraduate degree programs

to three years with the intention that students should be able to graduate at an earlier age than in the

past, in line with graduates from other European countries. On the other hand, the reform has

introduced (at least at the beginning) greater flexibility in the degree structure along with a wider

range of curricula offered to students. With more autonomy for each University, a territorial

decentralization process occurred with the proliferation of new small universities and/or of new

branches of historical universities.

More recently (2009 – 2013) a reversal of policy is taking place: many branches of the

historical universities are closing due to new requirements for university degree programs, in terms of

2 For a comprehensive English summary of the Italian university system, its evolution within the Bologna Process, and the

Almalaurea surveys main results see A. Cammelli, G. Antonelli, A. di Francia, G. Gasperoni, M. Sgarzi (2010).

3

numbers of students enrolled and stable teaching staff. Furthermore the role, and thus the existence, of

recently established small universities is put in doubt.

According to a document by CUN3, the Italian university system suffers from severe

emergencies:

(i) the percentage of GNP for higher educations is less than 1 %, compared with an OECD

average of 1,6 %. Italy ranks 32nd

, out of 37 countries;

(ii) the public financing of University (FFO, Fondo di Finanziamento Ordinario) decreased

in absolute inflation adjusted terms of about 20% since 2009;

(iii) a strong reduction in teaching staff occurred through a freezing of the turnover: five

teacher out for one teacher in;

(iv), even if graduates are only about 19% of the population aged 30-34 years (the European

average is 30%), the public financing for scholarships and other services for low income students has

been heavily cut in recent years;

(v) the total annual enrolment passed from 340.000 in 2003-4 to about 280.000 in 2011-12,

mostly because of the vanishing boom of new enrolment from the stock of older people, enhanced by

the 3+2 reform at the beginning of the century;

(vi) on the supply side, the numbers of courses of studies offered is steadily decreasing after

the peak of more than 5.500 in year 2007-08. Now (2012) they are about 4.3004.

This paper analyses the experience of a small Italian university, the Piemonte Orientale5,

assumed as a case study, in order to evaluate its role in human capital accumulation, a necessary, even

though not sufficient condition, for economic development. The crucial point is: does this small

university satisfie a specific demand otherwise not satisfied by larger existing universities? Actually,

without this new University, a relevant number of potential students could not enroll, attend and

graduate without huge economic cost for their families and strong personal effort.

The important role of small universities in the human capital accumulation in the geografichal

recruitment basin, is a phenomenon with medium and long term implications: small universities not

3 Consiglio Universitario Nazionale, National University Council, the official representative institution of all professors and

researchers. 4 CUN Consiglio Universitario Nazionale, Dichiarazione del Consiglio Universitario Nazionale per l’Università e la Ricerca

Gennaio 2013 “Le emergenze del sistema”, avallabile at www.cun.it 5 The University of Piemonte Orientale ”Amedeo Avogadro”, a spin-off of the University of Torino, became an independent

University in 1998. It is based in three medium sized main county (provincial) cities (Alessandria, Novara and Vercelli), has

(2012) seven Faculties and twelve Research Departments. Enrolled students are about 10,000 and teaching and

administrative staff, are each about 350 people.

4

only could ease the congestion that plagues larger universities but also perform a key role in the

development of peripheral territorial systems both on the demand and supply side.

The evaluation of the impact of small universities becomes urgent due to the recently

instituted university assessment process based on rankings6 (productivity, research, teaching,

professors’ CVs, international relations) and efficiency course requirements. The paper also analyzes

the public policy implications of recent regulation by the so-called Gelmini Reform. The new law and

its administrative implementation reduce the opportunities for a decentralized supply of academic

programs and therefore strengthen the role of student mobility.

The paper is organized as follows. Section 2 breathly describes tertiary education in Italy,

Section 3 analyzes the general context of student territorial and social mobility, Section 4 surveys the

earlier empirical evidence literature, Section 5 provides information about the dataset employed in the

empirical analysis, with summary statistics, Section 6 analyzes the selected models, describes the

estimation procedures and comments on the empirical findings. Finally, Section 7 summarizes the

discussion and suggest some policy implications.

2. Tertiary Education in Italy

In this version (Almalaurea May 2013), we skip this paragraph which is intended to summarize the

long season of reform and counter reform of higher education in Italy from 1995 to 2013.

This process, long overdue, started with the Italian budget law for 1995, which granted each

University greater autonomy on expenditures decisions. The full reform (law 509/90) introduced,

according to the Bologna Process, the 3 + 2 system (bachelor + master degrees system). As soon as

the universities started to fully exploit their higher financial and organizational autonomy, the supply

of new courses of study boomed. In order to control this outcome, a policy reversal occurred in 2005

with the introduction of “Minimun Requirements” for teaching staff and other characteristics. A long

series of budget cuts were introduced in the following years, both because of macroeconomic

constraints for the public expenditure and in order to control the behaviour of single universities.

This progressive counter reform has been consistently arranged in a sistematic way in the so called

6 Censis, for example, provides Italian universities rankings. Recently the MIUR and the National Evaluation Committee

(now ANVUR) started gathering data on research and teaching in order to evaluate the university system.

5

Riforma Gelmini which, under the cover of efficiency, rationalization and better governance and

organization, actually produced an increasing centralization and rigidity of the system.

3. Territorial and Social Mobility in Italy: an Overview.

An interesting picture of the Italian University System (as it appears from the set of Academic

Institutions belonging to the AlmaLaurea Consortium) can be obtained simply analyzing two

indicators: (i), the percentage of local graduates (living in the same county (provincia) where the

university is based, and (ii), the percentage of graduates with at least one parent holding a university

degree. Plotting the normalized values of both indicators we get a cloud (Graph1), with a triangular

shape.

We associate each university with a different label: Urban Niche Universities7, (specializing

most often in peculiar curricula, such as gastronomique science in Bra), Historical Universities8

mainly based in regional main city and the New Small Universities9, the picture shows a surprising

significance. We can identify three well defined horizontal belts from the top educated families to the

bottom. The top belt includes all Urban Niche Universities, the middle belt includes all Historical

Universities with some outliers. The bottom belt includes all the New Small Universities. In each belt

we find institutions with a higher or lower value of the local graduates indicator, which is consistent

with the perceived quality of the university (the lower, the better).

Table 1 Social and Territorial Mobility Indicators for Graduates in Almalaurea Universities (average 2003-2011)

7 The thirteen Universities considered are: LIUC Castellanza, LUM Casamassima, Milano IULM, Milano San Raffaele,

Perugia Stranieri, Roma Campus Bio Medico, Roma Foro Italico, Roma LUMSA, Roma LUSPIO, Roma TRE, Scienze

Gastronomiche Bra, Siena Stranieri, Venezia IUAV 8 The twenty-six Universities considered are: Bari, Bari Politecnico, Bologna, Cagliari, Camerino, Catania, Ferrara, Firenze,

Genova, Marche Politecnica, Messina, Modena e Reggio Emilia, Napoli Federico II, Napoli L'Orientale, Napoli Parthenope,

Padova, Parma, Perugia, Roma La Sapienza, Sassari, Siena, Torino, Torino Politecnico, Trieste, Urbino, Venezia Ca' Foscari. 9 The twenty-two Universities considered are: Basilicata, Bolzano, Calabria, Cassino e Lazio Meridionale, Catanzaro Magna

Grecia, Chieti e Pescara, Foggia, Insubria, L'Aquila, Molise, Napoli Seconda Università, Piemonte Orientale, Reggio

Calabria Mediterranea, Salento (Lecce), Salerno, Sannio, Teramo, Trento, Tuscia, Udine, Valle D'Aosta, Verona.

6

University

Absolute Value Normalized Value

Local

Graduates

(%)

Parent(s) Holding

University Degree (%)

Local

Graduates

Parent(s) holding

University Degree

Scienze Gastronomiche Bra 4,70 40,70 0,00 0,70

Venezia IUAV 22,83 27,63 0,23 0,41

Chieti Pescara 27,10 16,53 0,28 0,16

Siena Stranieri 28,90 29,10 0,30 0,44

Padova 31,03 20,38 0,33 0,25

Bologna 31,13 30,35 0,33 0,47

Camerino 33,80 21,08 0,36 0,26

Perugia Stranieri 34,05 25,15 0,37 0,35

Trieste 34,28 26,38 0,37 0,38

Parma 34,70 25,50 0,37 0,36

Tuscia 35,13 14,33 0,38 0,11

Urbino 35,30 21,10 0,38 0,26

Venezia Ca' Foscari 35,50 21,55 0,38 0,27

L'Aquila 36,85 19,85 0,40 0,23

IULM Milano 37,65 30,80 0,41 0,48

Siena 37,85 23,33 0,41 0,31

LIUC Castellanza 38,07 35,53 0,42 0,59

Ferrara 42,33 22,03 0,47 0,28

Firenze 44,33 25,50 0,49 0,36

Roma LUSPIO 44,50 36,70 0,50 0,61

San Raffaele Milano 45,30 43,93 0,51 0,78

Teramo 45,55 18,45 0,51 0,20

Marche Politecnica 48,90 21,50 0,55 0,27

Roma Campus Bio-Medico 49,30 53,83 0,56 1,00

Molise 49,58 13,95 0,56 0,10

Piemonte Orientale 49,68 13,45 0,56 0,09

Udine 49,85 17,23 0,56 0,18

All Almalaurea Universities 53,47 23,40 0,61 0,31

Modena e Reggio Emilia 54,53 22,23 0,62 0,29

Cagliari 55,03 17,07 0,63 0,17

7

University

Absolute Value Normalized Value

Local

Graduates

(%)

Parent(s) Holding

University Degree (%)

Local

Graduates

Parent(s) holding

University Degree

Perugia 55,18 24,08 0,63 0,33

Cassino e Lazio Meridionale 56,65 13,88 0,65 0,10

Torino Politecnico 57,68 30,50 0,66 0,47

Calabria 57,93 18,00 0,66 0,19

Sannio 59,10 22,75 0,68 0,30

Roma La Sapienza 59,43 28,78 0,68 0,44

Catania 59,65 23,78 0,69 0,32

Napoli L'Orientale 60,30 26,30 0,69 0,38

Roma LUMSA 60,60 38,80 0,70 0,66

Trento 60,93 20,85 0,70 0,26

Verona 61,75 15,95 0,71 0,15

Salerno 61,80 22,85 0,71 0,30

LUM Casamassima 61,90 41,30 0,71 0,72

Genova 65,20 28,28 0,75 0,42

Lecce Salento 65,85 14,35 0,76 0,11

Sassari 66,73 12,23 0,77 0,06

Napoli Seconda Università 67,23 19,93 0,78 0,24

Catanzaro Magna Grecia 68,30 9,43 0,79 0,00

Roma Foro Italico 68,40 23,53 0,79 0,32

Messina 68,68 21,33 0,80 0,27

Insubria 71,10 13,20 0,83 0,08

Bari Politecnico 71,60 26,20 0,83 0,38

Bolzano 72,33 20,90 0,84 0,26

Torino 72,55 21,78 0,85 0,28

Bari 75,60 16,18 0,88 0,15

Napoli Federico II 75,70 27,10 0,89 0,40

Reggio Calabria Mediterranea 77,23 27,38 0,90 0,40

Basilicata 77,75 13,90 0,91 0,10

Roma TRE 78,68 34,95 0,92 0,57

Valle D'Aosta 80,83 12,33 0,95 0,07

Napoli Parthenope 81,10 20,80 0,95 0,26

Foggia 84,90 14,80 1,00 0,12

8

Graph 1 Dispersion Diagram by Territorial and Social Mobility: the Position of the Urban Niche Universities

9

Graph 2 Dispersion Diagram by Territorial and Social Mobility: the Position of the Historical Universities

10

Graph 3 Dispersion Diagram by Territorial and Social Mobility: the Position of the New Small Universities

11

4. The Empirical Evidence Literature

The empirical literature about territorial mobility of university students or graduates is very

limited.

Cappellari and Lucifora (2008) estimate the impact of the italian reform on the decision to go

to college: they find that individuals leaving secondary school, after the reform have a probability of

going to college that is 10 % higher, compared to individuals making the choice under the old system.

This increase is concentrated among individuals with good high-school performance and low parental

educational background. This result may be interpreted as an indication of the existence of constraints

(pre-reform) – for good students from less affluent households – on the optimal schooling decision.

Rotaris, Danielis and Rosato (2012) study the choice between commuting and staying in the

university town by university students in the Trieste area. Using a theoretical model derived from

urban and regional science, they find that the crucial indifference distance between living in the

university or home town is about 65-75km. This result may be assumed to imply that the household

economic constraints prevent students to enroll in the university if there is none within the above

range.

According to Ciriaci and Nuzzi (2012, p. 376) the marginal effect of the research quality on

the probability to migrate for the enrolment in the University is equal to 0,69 (an increase of 1% of the

avarege value of the University rating implied an increase of the probability of migrating of 0,69%).

The higher the education level of the parents, the higher the probability of migrating.

Hertz et al. (2001) focus on the intergenerational persistence of educational inequality. The

paper analyzes 42 countries according to educational persistence across generations. Empirical

evidence shows significant differences among countries, in particular between Latin America and

European Nordic countries. The first group is characterized by a higher intergenerational persistence

of educational differences, the second exhibits a low serial correlation. In Italy the significant

correlation is further evidence to support the claim that “weak” students (on account of their social

economic and cultural background) find it more difficult to graduate. For this reason, it is important to

analyze the demand satisfied by small universities to check if they play a specific educational role.

According to Dotti, Fratesi, Lenzi, Percoco (2010) and Ciriaci (2005), Coniglio and E. Prota

(2008), Fratesi e Percoco (2010), migrant graduates from southern Italy who obtained their degree in

northern Italy show a low probability of returning home. A gravitational model is used to analyse the

12

students mobility flows, with the result that a good University and a dynamic labour market are the

main drivers. In particular long distance mobility depends on labour market conditions, while short

distance mobility appears linked to the University reputation.

Aina C. and Baici E. (2012) analyse the student choice between to stay or to move. In

particular, the paper analyses the effects of the number of university degree courses available in the

province of residence on the probability of studying in that province rather than moving to a different

one. The supply of degree courses outside the province of residence is weighted by a spatial matrix

where the distance between the province of residence and any other province is imputed in minutes.

The results confirm that the probability of enrolling in a universities located in the same province of

residence is positively correlated to the number of courses available. In particular, the probability of

enrolling at university in the province of residence is higher for the departments where the attendance

is compulsory (i.e. Engineering). The enrolment behaviour of freshmen differs according to gender,

with females more sensitive to the supply of local degree courses.

5. The data

The data on individual socio-economic conditions of the Piedmont graduates are drawn from

the Almalaurea survey, for 2008 and 2010.

Almalaurea (www.almalaurea.it) is a consortium of Italian universities with the support of the

Ministry for Education, University and Research. Almalaurea covers 77% of Italian graduates and, for

the year 2010, the total number of personal CV for graduates in 56 Italian universities was about 1,47

million. The Almalaurea database includes many quantitative and qualitative social and economic

variables on Italian graduates, based on answers to a questionnaire completed by all graduates from

AlmaLaurea universities.

The survey includes information on Italian graduates which are gathered to the eve of the

degree. The survey covers the 95% of the graduates. The sample data used in this work are two cross

section. The cross section related to 2008 contains 949010

individual observations: students awarded a

10 We dropped the observations for which the students have no choice (unique curricula or faculties) because the curricula

chosen is not offered by the Piemonte Orientale (for example veterinary or engineering), thus our sample contains 8703

individual observations.

13

bachelor’s degree11

in 2008 (regardless of the year of first enrolment) living in Piedmont (i.e. those

who, answering the questionnaire, give as current residence a town in Piedmont). The cross section

for 2010 contains 8737 individual observations. Even though the samples (2008 and 2010) do not

include data on Pavia and Milan Universities (also potential and actual choices according to MIUR

data), the samples are representative respectively of 87.4% and of 88.7% of Piedmont graduates.

In Table 2 we compare the samples (2008 and 2010) with MIUR data on Piedmont graduates

in order to consider the fraction of Piedmont students who graduated in non Almalaurea universities

(the flag Affiliated to Almalaurea consortium takes the values “1” for the universities recently

associated to Almalaurea, “2” for Almalaurea universities at the time of the survey, in 2008 and 2010

respectively, and “ blank” for those currently (2012) not associated to Almalaurea).

11Among graduates there are three types of students due to Italian university reform (law n. 509/1999): those who enrolled

before 2001 in an old system degree program and who graduated within the same program, those enrolled before 2001 in an

old system degree program who decided to switch to a “new” degree program and who therefore graduated in a 3 years

program, and those who enrolled after 2001 in a 3 years program.

14

15

Table 2: Graduates by Degree in all Universities and Almalaurea Universities (2010 and 2008)

2010 2008

Affiliated to Almalaurea

Consortium

(at 2010)

Universities TYPOLOGY OF DEGREE

Total

Affiliated to Almalaurea

Consortium

(at 2008)

Universities TYPOLOGY OF DEGREE

Total CDL L LS/LM

LSCU

/LMCU CDL L LS

LSCU

/LMCU

2 Aosta 1 10 2 0 13 2 Aosta 1 6 3 0 10

2 Bari 2 2 1 0 5 2 Bari 3 1 5 0 9

Bergamo 0 2 0 0 2 Bergamo 1 6 0 0 7

2 Bologna 5 45 30 3 83 2 Bologna 8 64 29 3 104

2 Bolzano 0 1 0 0 1 2 Bolzano 0 0 1 0 1

1 Bra - Scienze Gastronomiche

0 7 6 0 13 Bra - Scienze Gastronomiche

0 17 0 0 17

Brescia 0 1 0 2 3

2 Cagliari 0 1 1 0 2 2 Cagliari 0 1 2 1 4

2 Calabria 1 1 0 0 2 2 Calabria 3 2 0 0 5

2 Camerino 0 1 0 0 1 2 Camerino 2 4 0 0 6

2 Casamassima -

J.Monnet 0 1 0 0 1 Casamassima -

J.Monnet 0 3 0 0 3

2 Cassino 0 0 1 0 1 2 Cassino 0 1 0 0 1

2 Castellanza

LIUC 0 20 16 4 40 2 Castellanza

LIUC 1 13 13 5 32

2 Catania 1 1 0 0 2 2 Catania 1 2 2 1 6

2 Catanzaro 0 2 1 0 3 2 Catanzaro 1 1 0 1 3

2 Chieti e Pescara 0 30 30 0 60 2 Chieti e Pescara 0 121 15 0 136

1 Enna - KORE 0 0 0 3 3 Enna - KORE 0 2 0 0 2

2 Ferrara 0 4 2 1 7 2 Ferrara 1 4 4 1 10

2 Firenze 1 53 13 0 67 2 Firenze 0 24 14 0 38

2 Foggia 0 2 0 0 2 2

2 Genova 27 318 150 52 547 2 Genova 59 270 99 38 467

2 Insubria 0 17 9 3 29 Insubria 0 19 5 4 28

2 L'Aquila 0 4 0 0 4 2 L'Aquila 0 26 3 0 29

16

Affiliated to

Almalaurea

Consortium (at 2010)

Universities TYPOLOGY

Total

Affiliated to

Almalaurea

Consortium (at 2008)

Universities TYPOLOGY OF DEGREE

Total CDL L LS/LM LSCU/LMCU CDL L LS/LM LSCU/LMCU

1 Macerata 0 5 1 0 6 Macerata 2 7 0 0 9

Marche 0 2 1 0 3

2 Messina 1 0 0 0 1 2 Messina 36 30 15 4 85

Milano 23 218 134 54 429 Milano 40 225 113 21 399

Milano Bicocca 12 81 42 5 140 Milano Bicocca 17 96 37 0 150

Milano Bocconi 2 95 64 0 161 Milano Bocconi 6 108 102 0 216

Milano

Cattolica 17 156 110 21 304

Milano

Cattolica 30 162 102 7 301

2 Milano IULM 1 46 13 0 60 2 Milano IULM 3 58 20 0 81

Milano Politecnico

4 182 105 3 294 Milano Politecnico

14 180 107 0 301

2 Milano San

Raffaele 0 18 8 2 28 2

Milano San

Raffaele 0 20 9 7 36

2 Modena e

Reggio Emilia 1 10 3 2 16 2 Modena e

Reggio Emilia 0 5 2 0 7

2 Molise 0 0 0 1 1

1 Napoli Federico

II 2 4 3 1 10

Napoli Federico

II 1 1 3 0 5

2 Napoli II 1 13 6 2 22 2 Napoli II 6 7 3 1 17

2 Napoli L'Orientale

0 0 1 0 1 Napoli L'Orientale

0 2 0 0 2

1 Napoli

Parthenope 0 1 0 0 1

Napoli Pegaso 0 4 0 4 8

Novedrate - e-

Campus 0 13 0 0 13

2 Padova 2 27 11 0 40 2 Padova 4 42 15 0 61

Palermo 0 2 0 0 2 Palermo 1 2 0 0 3

2 Parma 1 13 7 4 25 2 Parma 11 17 15 8

51

17

Affiliated to

Almalaurea Consortium

(at 2010)

Universities TYPOLOGY

Total

Affiliated to

Almalaurea Consortium

(at 2008)

Universities TYPOLOGY OF DEGREE

Total CDL L LS/LM LSCU/LMCU CDL L LS/LM LSCU/LMCU

Pavia 19 254 169 85 527 Pavia 41 277 181 67 566

2 Perugia 1 3 3 1 8 2 Perugia 1 7 1 1 10

2 Perugia

Stranieri 0 7 1 0 8 2

Perugia

Stranieri 0 4 1 0 5

2 Piemonte

Orientale 31 983 210 157 1.381 2 Piemonte

Orientale 95 1027 219 107 1449

Pisa 0 17 20 4 41 Pisa 2 9 32 0 43

2 Reggio

Calabria 1 1 0 0 2 2

Reggio

Calabria 1 0 0 0 1

Roma

Mercatorum 0 4 0 0 4

2 Roma La Sapienza

3 13 8 1 25 2 Roma La Sapienza

4 12 7 0 23

Roma LUISS 0 3 0 1 4 Roma LUISS 1 1 1 0 3

2 Roma LUMSA 0 2 1 0 3 2 Roma LUMSA 0 4 1 0 5

Roma Marconi 0 37 6 8 51 Roma Marconi 0 31 1 0 32

2 Roma San Pio

V 0 1 5 0 6

Roma San Pio

V 0 2 1 0 3

Roma

TEL.M.A. 0 3 1 0 4 Roma

TEL.M.A. 0 6 1 0 7

1 Roma Tor Vergata

0 5 8 0 13 Roma Tor Vergata

1 7 10 0 18

2 Roma Tre 0 4 3 0 7 2 Roma Tre 1 1 3 1 6

Roma

UNINETTUNO 0 2 0 0 2 Roma

UNINETTUNO 0 1 0 0 1

Roma UNISU 0 3 1 5 9 Roma UNISU 0 1 0 1 2

2 Salento 1 2 0 0 3 2 Salento 0 1 1 0 2

2 Salerno 1 1 0 0 2

2 Sannio 0 1 0 0 1

18

Affiliated to

Almalaurea Consortium

(at 2010)

Universities TYPOLOGY

Total

Affiliated to

Almalaurea Consortium

(at 2008)

Universities TYPOLOGY OF DEGREE

Total CDL L LS/LM LSCU/LMCU CDL L LS/LM LSCU/LMCU

2 Sassari 0 0 0 1 1 2 Sassari 0 1 0 1 2

2 Siena 0 24 5 0 29 2 Siena 1 101 6 1 109

2 Siena Stranieri 0 0 1 0 1

2 Teramo 0 1 6 0 7 Teramo 0 1 4 0 5

2 Torino 476 5.875 2.685 752 9.788 2 Torino 820 5769 2423 551 9564

2 Torino

Politecnico 83 1.759 1.279 0 3.121 2

Torino

Politecnico 188 1776 1184 0 3154

Torrevecchia

Teatina - Leonardo da

Vinci

0 4 0 0 4 Torrevecchia

Teatina - Leonardo da

Vinci

0 2 0 0 2

2 Trento 0 1 4 0 5 2 Trento 2 3 0 0 5

2 Trieste 2 12 8 1 23 2 Trieste 3 11 5 1 20

2 Tuscia 1 12 0 0 13 2 Tuscia 1 36 2 0 39

2 Udine 0 0 2 0 2 2 Udine 0 1 2 0 3

1 Urbino Carlo

Bo 3 24 8 4 39

Urbino Carlo

Bo 8 25 3 1 37

2 Venezia Ca'

Foscari 1 22 12 0 35 2 Venezia Cà

Foscari 1 12 10 0 23

2 Venezia Iuav 0 2 9 0 11 2 Venezia Iuav 1 0 7 0 8

2 Verona 0 2 6 0 8 2 Verona 0 3 3 0 6

TOTAL 727 10.493 5.231 1.187 17.638 TOTAL 1.425 10.685 4.848 834 17.801

2010 2008

Piedmont

Graduates

TYPOLOGY OF DEGREE Piedmont

Graduates

TYPOLOGY OF DEGREE

CDL L LS/LM LSCU/LMCU Total CDL L LS/LM LSCU/LMCU Total

Piedmont in

AlmaLaurea

2008: absolute

value

645 9.366 4.553 987 15.551 Piedmont in AlmaLaurea

2008:

absolute value

1.260 9.490 4.144 733 15.627

percent 88,72% 89,26% 87,04% 83,15% 88,17% percent 88,42% 88,82% 85,48% 87,89% 87,79%

Piedmont

absent from AlmaLaurea

2010

82 1127 678 200 2087 Piedmont

absent from AlmaLaurea

2010

165 1195 704 101 2174

19

Source Almalaurea and Miur

LEGENDA

L = Laurea (First Cycle Degree/Bachelor - 180 ECTS);

LS/LM = Laurea Specialistica/Magistrale (Second Cycle Degree/Two years Master - 120 ECTS);

LSCU/LMCU = Laurea Specialistica/Magistrale a ciclo unico (Single Cycle Degree/Combined Bachelor and Master - 300/360

ECTS).

20

The dataset includes variables which characterize socio-economic background (proxied by

parents’ education, parents’ professional status); education (secondary school typology; secondary

school graduation marks); gender; student performance and student mobility (measured by the

distance between Faculty location and student home address in both kilometers and minutes); working

student status. Table 3 shows summary statistics of the main variables12

.

Table 3: Summary Statistics of the Sample Variables

2008 2010

Variable Description of variable Mean Std.

Dev. Min Max

Mean

Std.

Dev. Min Max

Marks average mean of the exam marks 26,19 2,20 19,64 30,00 26,18 2,14 19,53 30,00

Secondary

School leaving

certificate

a categorical variable which

takes values of 0 for

university-oriented

secondary schools

(specializing in classical

studies, science, art, modern

languages) and

progressively higher values

for technical or vocational

schools.

3,61 2,66 - 13,00 3,58 2,60 - 13,00

Secondary

School leaving

certificate marks

/100

Secondary School leaving

certificate marks 82,32 12,63 60,00 100,00 82,46 12,57 60,00 100,00

Duration duration of university

enrolment (years) 4,34 2,85 1,31 38,95 4,44 2,79 1,58 37,10

Erasmus

Dummy variable :

1 = the student took part in

Erasmus program

0,05 0,22 - 1,00 0,05 0,22 - 1,00

Age average average age at graduation

(years) 26,59 5,86 20,68 71,79 26,30 5,57 20,19 71,71

12 Most Almalaurea categorical variables are redefined by grouping together the “similar” categories in order to focus on the

effects of very different socio-economic background (low, medium and high) and the main typologies of secondary school

(university-oriented schools vs. technical or vocational schools) on university choice.

21

2008 2010

Variable Description of variable Mean Std.

Dev. Min Max

Mean

Std.

Dev. Min Max

Enrolment

difference between actual

age at enrolment and

standard enrolment age (19

years)

2,13 5,40 - 2,00 50,00 1,74 4,86 - 3,00 47,00

Mother

education

a categorical regressor:

1 = no certified school

attendance;

2 = primary school

certificate;

3 = lower secondary school

certificate;

4 = upper secondary school

certificate;

5 = university degree.

3,54 1,04 - 5,00 0,73 0,71 - 2,00

Father education same 3,58 1,09 - 5,00 0,75 0,74 - 2,00

Degree final

mark /110

degree mark (the laude is

accounted as 3 points) 102,00 8,61 74,00 113,00 101,69 8,93 73,00 113,00

Working Student

status

Dummy variable:

1 = studing worker 0,12 0,32 - 1,00 2,12 0,57 1,00 4,00

Dummy Mother

holding a Un.

degree

Dummy variable:

1 = holding a university

degree

0,15 0,36 - 1,00 0,15 0,36 - 1,00

Dummy Father

holding a Un.

degree

same 0,18 0,39 - 1,00 0,18 0,39 - 1,00

22

2008 2010

Variable Description of variable Mean Std.

Dev. Min Max

Mean

Std.

Dev. Min Max

Mother

professional

status

a categorical variable:

1 = self-employed, family

co-worker, partner of

worker cooperative);

2 = entrepreneur;

3 = professional;

4 = senior executive,

manager;

5 = cadre;

6 = white collar worker;

7 = blue collar worker;

9 = nonworking.

5,92 2,33 1,00 9,00 5,89 2,28 1,00 9,00

Father

professional

status

same 4,58 2,12 1,00 9,00 4,57 2,13 1,00 9,00

Gender Dummy variable:

1 = male 0,38 0,49 - 1,00 0,38 0,49 - 1,00

Distance minutes

distance between student

address and university

location in minutes

45,78 98,66 - 1467 37,70 53,78 - 987,00

Dummy of

overlapping

Dummy variable:

1 = the province of the

university is the same of the

residence

0,24 0,43 - 1,00

Distance km

distance between student

address and university

location in kilometers

38,88 61,15 - 983,00 33,13 88,97 - 1.585,00

23

6. The Model and Empirical Results

The probability for a Piedmont student to choose and graduate in Piemonte Orientale instead

of other universities (which offer same or similar programs) is modeled by a probit/logit:

(1)

(2)

(3)

where the observed values of yi,j are outcomes for individual ‘i’ enrolled in faculty ‘j’

generated by the regressors. X is a vector of exogenous variables representing: individual’s personal

characteristics (such as gender and age), pre-university qualifications (such as mark at Secondary

school leaving certificate and type of school attended13

), indicators of family background (for example

income proxied by parents’ employment typologies), possible peer group effects, and distances to

university measured by three variables (for 2008 estimates the inverse of distance, the time necessary

to reach the Faculty for 2010 estimates a dummy variable which takes value of one if student address

and faculty location belong to the same province and the time necessary to reach the Faculty

analogously to 200814

). β is a set of parameters to be estimated and ε is the usual white noise error

term. The model is a binomial probit/logit for the individual's probability to choose Piemonte

Orientale, where yi,j = 1 if the individual chooses it and yi,j = 0 otherwise. The Piemonte Orientale

locations are Alessandria, Novara, Vercelli, Acqui Terme, Alba-Bra, Asti, Biella and Casale

Monferrato for 2008 while in 2010 Acqui Terme and Biella are no longer active.

We estimate four main models: two different models, where parent occupation and parent

education are alternatively entered as additional variables into the model (due to huge correlation),

and two distributions in order to model the probability function (logit versus probit). Parent education

is proxied by either categorical variables (Mother’s education and Father’s education) or dummy

variables (Dummy Mother holding a Un. degree and Dummy Father holding a Un. degree). This

generates two further sub-models, called “Model A bis” for logit or probit distribution respectively

(Probit model A_bis and Logit model A_bis).

13 Further development of this work should consider also the endogeneity problem which could arise if secondary school

leaving certificate is chosen to credibly signal some information relevant for the labor market. Thus we are gathering

information about neighborhood secondary school leaving certificate for each province in order to control for it. 14 The change in one variable representing the distance permit to overtake the 2008 estimates inconvenient, that is the

marginal effect greater than one.

24

In all models, the main explanatory variable estimates have all turned out to be significant

and their signs are as expected and stable.

Table 4 shows empirical results for 2008 and Table 6 for 2010. Table 5 (2008) and Table 7

(2009) contain the marginal effects of each independent variable, that is the derivative of the

prediction function, which, by default, is the probability of success following probit/logit, that is the

choice of Piemonte Orientale. By default, margins evaluate this derivative for each observation and

report the average of the marginal effects.

In particular, the educational and socio-economic background crucially affects university

choice. Piemonte Orientale is chosen with higher probability by technical or vocational secondary

school students. This evidence is confirmed by 2010 estimates. Further the Secondary School leaving

certificate mark is a significant explanatory variable, but its rule is ambiguous: for 2008 estimates, the

higher the mark, the lower is the probability of Piemonte Orientale choice; for 2010 estimates, the

effects became positive. This finding can be explained by the dropout of the students of the old Italian

University system (the so called “vecchio ordinamento”) or by an improvement in the relative

positioning of the Piemonte Orientale or by both causes. An experience of study abroad (Erasmus)

negatively affects the probability of Piemonte Orientale choice, but this results is not confirmed by the

estimates on 2010. Parents education, whether proxied by graduate dummies or categorical variables,

negatively affects the small university choice. In particular, if his/her parents are graduates then the

student will more likely enrol and graduate in the historical universities which his/her parents had

chosen. Predictably, parents’ occupations affect university choice through the student’s economic

opportunities.

The small university is chosen with higher probability by “weaker” students: these would be

female, living in peripheral municipalities suffering from structural problems, with lower average

marks (which is a signal of either lower abilities or stronger selection).

The Working Student status plays an unexpected negative role, significant in 2008’

estimates, that is, if the student works then he/she has a smaller probability to choose Piemonte

Orientale. This finding can be explained as the net effect of opposite forces: if the studying workers

usually do not attend classes their choice is independent of distance so they choose the university

where they work (very likely the bigger city) and not the one where they live; studying worker choose

a local university only because it allows them to effectively keep their job and attend classes. Another

explanation focuses on the higher concentration of good jobs in bigger cities where the historical

universities are located: this is where, according to the gravitational model of regional economics,

25

government, banking, insurance and financial services, research activities and company headquarters

also tend to be located. The empirical results for 2010 does not confirm this evidence: the Working

Student status is always negative in all the model estimates but is not significant, at usual confidence

levels.

Duration is not significant in all the models. This evidence supports the idea that the

educational supply of small universities is comparable to that offered by the bigger universities.

The marginal effects for these explanatory variables are significant, sign coherent, stable and

modest in all models, except the distance factor.

Actually, distance, and therefore student mobility, does matter. The role of student mobility

in the Italian university system is crucial: the mobility of student is strictly constrained by the

accessibility of the supply point. The very low mobility of university students in Italy is mainly due to

poor and unequal availability of low cost student accommodation, expensive and inefficient

commuting opportunities, and finally to social, economic and cultural constraints.

The marginal effect of distance is always significant and very large. In 2010 estimates the

marginal effect is 13% and in 2008 estimates is greater than one15

.

Our empirical results show that Piemonte Orientale, a small university satisfies a particular

tertiary education demand. Also its demographic recruitment basin is characterised by a specific

geographical area.

15 This seems like a strange result considering that the values of the probability function are bound between 0 and 1.

Nevertheless this can be explained by the computation technique of marginal effects. The marginal effect is the derivative,

that is the approximate change in the dependent variable y for a one-unit change in a regressor x. Because y is between 0 and

1, the change in y obviously cannot be greater than 1, but the marginal effect computes the approximate change. The

derivative at a point is the slope of the tangent line to the curve at that point. Thus the slope of the tangent line, at the point

Inverse_distance_km is greater than 1, therefore distance does matter.

26

27

Table 4 Model Estimations 2008

Variable Probit model A Probit model A_bis Probit model B Logit model A Logit model A_bis Logit model B

Secondary School leaving

certificate 0.1930736 *** 0.1877444 *** 0.1986058 *** 0.3514739 *** 0.3400375 *** 0.3631498 ***

Secondary School leaving

certificate marks /100 -0.0063332 *** -0.0060648 *** -0.0063148 *** -0.0119645 *** -0.01145 *** -0.0119261 ***

Erasmus -0.2898485 *** -0.3043228 *** -0.3497474 *** -0.5763653 *** -0.6017015 *** -0.6990103 ***

Parents education

Mother education -0.0204652 -0.0212551

Father education -0.0993238 *** -0.1851035 ***

Dummy Mother holding a

Un. degree -0.1481552 *** -0.2780425 **

Dummy Father holding a

Un. degree -0.1853156 ** -0.3376159 ***

Working Student status -0.1659365 *** -0.1538005 *** -0.1527958 ** -0.3066475 *** -0.2806492 *** -0.2804307 **

Mother professional status -0.0320895 ** -0.0553512 **

Father professional status -0.0178927 . -0.0278286

Gender (1=male) -0.1531113 *** -0.1594376 *** -0.1512158 *** -0.271704 *** -0.2813243 *** -0.2666229 ***

Marks average -0.0401973 *** -0.0399961 *** -0.0385565 *** -0.0730546 *** -0.0719707 *** -0.0695823 ***

Duration -0.0022652 -0.001577 -0.0000456 -0.0077752 -0.0066242 -0.0040002

Inverse_ Distance _km 66.11107 *** 66.3036 *** 66.08145 *** 119.0929 *** 119.3543 *** 119.0725 ***

Distance _minutes -0.0121669 *** -0.0121238 *** -0.0118654 *** -0.0230637 *** -0.0229418 *** -0.0225088 ***

Constant 0.3126587 0.2543957 0.2638964 0.7351999 0.6203103 0.6173739

Log likelihood -2833.603 -2912.779 -2770.9677 -2819.1214 -2898.3688 -2756.6194

LR χ2

1112.640 1146.010 1073.73 1141.61 1174.83 1102.42

28

Prob > χ2 0.000 0.000 0.000 0.000 0.000 0.000

Goodness of Fit

Pseudo R2

0.164 0.1644 0.1623 0.1684 0.1685 0.167

Efron's R2 0.197 0.1960 0.1950 0.201 0.2 0.199

Cragg & Uhler's R2

0.231 0.2310 0.2280 0.237 0.237 0.234

McKelvey&Zavoina's R2

0.364 0.3760 0.3570 0.372 0.385 0.365

McFadden's Adj R2

0.161 0.1610 0.1590 0.165 0.165 0.163

Correctly classified 86.880% 86.930% 87.150% 86.88% 86.96% 87.13%

The similar categories of the categorical variables have been grouped in order to reduce the categories.

Signif. codes Pr(>|t|): 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 .. if Pr(>|t|) around 0.1

29

Table 5 Marginal Effects 2008

Variable

Marginal Effects

Probit model A Probit model A_bis Probit model B Logit model A Logit model A_bis Logit model B

Secondary School leaving certificate 0.04063 *** 0.039155 *** 0.041592 *** 0.037705 *** 0.036078 *** 0.038726 ***

Secondary School leaving certificate marks /100 -0.00133 *** -0.00126 *** -0.00132 *** -0.00128 *** -0.00121 *** -0.00127 ***

Erasmus -0.05231 *** -0.05394 *** -0.06073 *** -0.05083 *** -0.052 *** -0.05878 ***

Parents education

Mother education -0.00431 -0.00228

Father education -0.0209 *** -0.01986 ***

Dummy Mother holding a Un. degree -0.02912 ** -0.02744 **

Dummy Father holding a Un. degree -0.03615 *** -0.0331 ***

Working Student status -0.03244 ** -0.02996 *** -0.02989 *** -0.0301 *** -0.02745 ** -0.02756 **

Mother professional status -0.00672 *** -0.0059 **

Father professional status -0.00375 -0.00297

Gender (1=male) -0.03156 *** -0.03255 *** -0.03104 *** -0.02847 *** -0.02914 *** -0.02779 ***

Marks average -0.00846 *** -0.00834 *** -0.00807 *** -0.00784 *** -0.00764 *** -0.00742 ***

Duration -0.00048 -0.00033 -9.55E-06 -0.00083 -0.0007 -0.00043

Inverse_ Distance _km 13.9124 *** 13.82805 *** 13.83878 *** 12.776 *** 12.66343 *** 12.6979 ***

Distance _minutes -0.00256 *** -0.00253 *** -0.00248 *** -0.00247 *** -0.00243 *** -0.0024 ***

30

Table 6 Model Estimations 2010

Variable Probit model A Probit model A_bis Probit model B Logit model A Logit model A_bis Logit model B

Secondary School leaving

certificate 0,1300678 *** 0,1056132 *** 0,0894115 ** 0,2345867 *** 0,19893 ** 0,1805844 **

Secondary School leaving

certificate marks /100 0,033846 *** 0,0341418 *** 0,0337214 *** 0,0679716 *** 0,068234 *** 0,0686103 ***

Erasmus

Dropped for

collinearity

Dropped for

collinearity

Dropped for

collinearity

Dropped for

collinearity

Dropped for

collinearity

Dropped for

collinearity

Parents education

Mother education -0,0845634 ** -0,088066 **

Father education 0,0930405 0,1456772

Dummy Mother holding a

Un. degree -0,0983893 ** -0,2755668 **

Dummy Father holding a

Un. degree -0,0191003 -0,0278949

Working Student status -0,0767364 -0,0347461 -0,1041622 -0,2180589 -0,1780359 -0,1517538

Mother professional status -0,1940208 *** -0,338123 ***

Father professional status 0,0200519 0,1003734

Gender (1=male) -0,477122 *** -0,440261 *** -0,4589813 *** -0,9737691 *** -0,9194377 *** -0,7736109 ***

Marks average -0,2110604 *** -0,2068851 *** -0,2034554 *** -0,4224529 *** -0,4191691 *** -0.3497079 ***

Duration 0,00415 -0,0001892 -0,011633 0,003168 -0,0035472 -0,0809503

Dummy of overlapping (same

province for the university and

for the residence) 1,050483 *** 1,026934 *** 1,072493 *** 2,369154 *** 2,380334 *** 2,5063 ***

Distance _minutes -0,0104464 *** -0,0111745 *** -0,011768 *** -0,0259758 *** -0,0278186 *** -0,0300284 ***

Constant 1,005449 1,10339 1,663208 2,657045 2,938163 4,3781

31

Log likelihood -91,051427 -91,498954 -88,86218 -89,385605 -89,486758 -86,79113

LR χ2

47,97 47,07 52,34 51,3 51,09 56,49

Prob > χ2 0 0 0 0 0 0

Goodness of Fit

Pseudo R2

0,2085 0,2046 0,2275 0,223 0,2221 0,2455 0,2221

Efron's R2 0,205 0,211 0,225 0,237 0,25 0,268

Cragg & Uhler's R2

0,257 0,252 0,279 0,273 0,272 0,299

McKelvey&Zavoina's R2

0,461 0,469 0,499 0,568 0,585 0,619

McFadden's Adj R2

0,208 0,205 0,228 0,223 0,222 0,246

Correctly classified 92,49% 92,49% 93,46% 93,46% 93,70% 94,19%

The similar categories of the categorical variables have been grouped in order to reduce the categories.

Signif. codes Pr(>|t|): 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 .. if Pr(>|t|) around 0.1

32

Table 7 Marginal Effects 2010

Variable

Marginal Effects

Probit model A Probit model A_bis Probit model B Logit model A Logit model A_bis Logit model B

Secondary School leaving certificate 0,0103732 *** 0,0084459 *** 0,0065337 *** 0,0071012 *** 0,0058522 *** 0,0041623 ***

Secondary School leaving certificate marks /100 0,0026993 *** 0,0027303 *** 0,0024642 *** 0,0020576 *** 0,0020073 *** 0,0018125 ***

Erasmus

Dropped for

collinearity

Dropped for

collinearity

Dropped for

collinearity

Dropped for

Collinearity

Dropped for

collinearity

Dropped for

collinearity

Parents education

Mother education -0,0067441 * -0,0026659 *

Father education 0,0074202 0,0044098

Dummy Mother holding a Un. degree -0,0076046 * -0,0077226 *

Dummy Father holding a Un. degree -0,0015171 ** -0,0008164 **

Working Student status -0,0061199 -0,0027786 -0,0076116 -0,0066009 -0,0052375 -0,0083183

Mother professional status -0,014178 *** -0,0096646 ***

Father professional status 0,0014653 -0,0012519

Gender (1=male) -0,0354383 *** -0,0329195 *** -0,0312698 *** -0,0275918 *** -0,0253686 *** -0,0221489 ***

Marks average -0,0168325 *** -0,0165445 *** -0,0148674 *** -0,0127881 *** -0,0123312 *** -0,0110833 ***

Duration 0,000331 -0,0000151 -0,0008501 0,0000959 -0,0001044 -0,0005724

Dummy of overlapping (same

province for the university

and for the residence) 0,1373634 *** 0,1331495 *** 0,1323324 *** 0,1427796 *** 0,1403953 *** 0,1392002 ***

Distance _minutes -0,0008331 *** -0,0008936 *** -0,0008599 *** -0,0007863 *** -0,0008184 *** -0,0007876 ***

33

34

7. Concluding Remarks and Implication for Public Policy

Over the last twenty years, the Italian university system underwent an important reform

process which was launched in the 1990s and is still going on with the implementation of a new

Reform Law (Act 240/2010). This latter is actually a Counter reform, a typical Italian specialty,

which is going to reverse the trend experienced in the previous 15 years.

Due to the need for greater institutional autonomy and self-regulation, the first set of reforms

affected the management model of the higher education system which moved away from external,

highly centralized control to more decentralized and internal control.

More recently a policy reversal has occurred, with a view to limit the excessive proliferation

of branches (and/or new universities) and curricula, and therefore to control the financial viability of

the system.

Within this framework, many controversial issues emerge. These include university autonomy,

efficiency and role of the recent small universities, affordability of higher education and

intergenerational mobility versus student mobility.

In this work we have provided a first analysis of the role that small universities play in human

capital accumulation, focusing on Piemonte Orientale as a case study. The model set up considers the

probability of choosing to enrol in Piemonte Orientale vs. other universities, given as explanatory

variable a set of social, economic and cultural variables. We have estimated four main binary outcome

models (plus two sub models for the parents’ education variables): two models where parent

occupation and parent education are alternatively entered as variables, and two distributions in order to

model the probability function (logit versus probit).

The econometric exercise carried out in this paper for 2008 and 2010 highlights the important

role played by small universities in both reducing the congestion in mega universities (and therefore

improving the efficiency of the educational process), and promoting the development of human

capital, and through this, contributing to economic growth and social mobility.

The most important findings of this paper can be summarized as follow:

(i) small universities satisfy a specific demand which cannot be satisfied by larger

universities, because of distance, socio-economic family background, educational background, gender,

35

parents education. This crucial role of the modest socioeconomic background and the low mobility of

the students are confirmed into 2010 estimates;

(ii) student mobility is strictly constrained by the accessibility of the supply point: the very

low mobility of university students in Italy is mainly due to poor and unequal availability of low cost

student accommodation, too expensive and inefficient commuting opportunities, and to social,

economic and cultural constraints;

(iii) family background, as measured by parental education, crucially affects the university

choice of the children, at least in the sense that a low indicator of parents education is the dominant

factor of the demand for higher education in small universities, which is mostly generated and not

“attracted”;

(iv) educational background is a choice driver, i.e. the small university is chosen with a higher

probability by vocational school students (not by university-oriented secondary school students). This

result is confirmed in 2010, while the poor performance in secondary school leaving certificate in

2010 does not increase the probability to choose Piemonte Orientale vs. mega or historical

universities, suggesting a evolution in recruitment;

(v) the small university is chosen with higher probability by “weaker” students (female, living

in peripheral small town).

Further research imply addressing the endogeneity problem which could arise if the secondary

school choice as a signal for future enrolment in university.

36

REFERENCES

Aina C. and Baici E. (2012), “To Stay or to Move? Freshmen and University Accessibility in Italy”, Rivista

Internazionale di Scienze Sociali, 1: 25-44

Almalaurea Survey, Profilo dei laureati – Condizione occupazionale dei laureati www.almalaurea.it

Cammelli A., Antonelli G., di Francia A., Gasperoni G., Sgarzi M. (2010), Employability and Mobility of

Bachelor Graduates in Italy: Mixed Outcomes of the Bologna Process, presentato a: International

Conference on “Employability and Mobility of Bachelor Graduates in Europe – Results of the Bologna

Process”, 30 September – 1 October 2010, Berlin, Germany

Cappellari L. and Lucifora C. (2008), “The Bologna Process and College Enrolment Decisions”, Labour

Economics, 16(6): 638–647.

Ciriaci D. and Nuzzi A. (2012), “Qualità dell’Università e mobilitàdei laureati italiani: evidenze empiriche e

proposte di policy”, Istituzioni del Federalismo Rivista di studi giuridici e politici N.2 2012 • ANNO

XXXIII - aprile/giugno.

Hertz T. (2001), “Education, Inequality and Economic Mobility in South Africa” Ph.D. diss., University of

Massachusetts.

Rotaris L., Danielis R. and Rosato P. (2012), “Scelta fra pendolarismo e domiciliazione degli studenti

universitari”, Scienze Regionali Italian Journal of Regional Science, 11(3): 51-74

Staffolani S. and Sterlacchini A. (2001), Istruzione universitaria, occupazione e reddito. Un’analisi empirica sui

laureati degli atenei marchigiani, Franco Angeli, Milan

Recent working papers

The complete list of working papers is can be found at http://polis.unipmn.it/index.php?cosa=ricerca,polis

*Economics Series **Political Theory and Law ε Al.Ex Series

Q Quaderni CIVIS

2013 n.205* Tiziana Caliman and Alberto Cassone: The choice to enrol in a small university: A case study of Piemonte Orientale

2013 n.204* Magnus Carlsson, Luca Fumarco and Dan-Olof Rooth: Artifactual evidence of discrimination in correspondence studies? A replication of the Neumark method

2013 n.203** Daniel Bosioc et. al. (DRASD): OPAL – Osservatorio per le autonomie locali N.2/2013

2013 n.202* Davide Ticchi, Thierry Verdier and Andrea Vindigni: Democracy, Dictatorship and the Cultural Transmission of Political Values

2013 n.201** Giovanni Boggero et. al. (DRASD): OPAL – Osservatorio per le autonomie locali N.1/2013

2013 n.200* Giovanna Garrone and Guido Ortona: The determinants of perceived overall security

2012 n.199* Gilles Saint-Paul, Davide Ticchi, Andrea Vindigni: A theory of political entrenchment

2012 n.198* Ugo Panizza and Andrea F. Presbitero: Public debt and economic growth: Is there a causal effect?

2012 n.197ε Matteo Migheli, Guido Ortona and Ferruccio Ponzano: Competition among parties and power: An empirical analysis

2012 n.196* Roberto Bombana and Carla Marchese: Designing Fees for Music Copyright Holders in Radio Services

2012 n.195* Roberto Ippoliti and Greta Falavigna: Pharmaceutical clinical research and regulation: an impact evaluation of public policy

2011 n.194* Elisa Rebessi: Diffusione dei luoghi di culto islamici e gestione delle conflittualità. La moschea di via Urbino a Torino come studio di caso

2011 n.193* Laura Priore: Il consumo di carne halal nei paesi europei: caratteristiche e trasformazioni in atto

2011 n.192** Maurilio Guasco: L'emergere di una coscienza civile e sociale negli anni dell'Unita' d'Italia

2011 n.191* Melania Verde and Magalì Fia: Le risorse finanziarie e cognitive del sistema universitario italiano. Uno sguardo d'insieme

2011 n.190ε Gianna Lotito, Matteo Migheli and Guido Ortona: Is cooperation instinctive? Evidence from the response times in a Public Goods Game

2011 n.189** Joerg Luther: Fundamental rights in Italy: Revised contributions 2009 for “Fundamental rights in Europe and Northern America” (DFG-Research A. Weber, Univers. Osnabrueck)

2011 n.188ε Gianna Lotito, Matteo Migheli and Guido Ortona: An experimental inquiry into the nature of relational goods

2011 n.187* Greta Falavigna and Roberto Ippoliti: Data Envelopment Analysis e sistemi sanitari regionali italiani

2011 n.186* Angela Fraschini: Saracco e i problemi finanziari del Regno d'Italia

2011 n.185* Davide La Torre, Simone Marsiglio, Fabio Privileggi: Fractals and self-similarity in economics: the case of a stochastic two-sector growth model

2011 n.184* Kristine Forslund, Lycia Lima and Ugo Panizza: The determinants of the composition of public debt in developing and emerging market countries

2011 n.183* Franco Amisano, Alberto Cassone and Carla Marchese: Trasporto pubblico locale e aree a domanda di mobilità debole in Provincia di Alessandria

2011 n.182* Piergiuseppe Fortunato and Ugo Panizza: Democracy, education and the quality of government

2011 n.181* Franco Amisano and Alberto Cassone: Economic sustainability of an alternative form of incentives to pharmaceutical innovation. The proposal of Thomas W. Pogge

2011 n.180* Cristina Elisa Orso: Microcredit and poverty. An overview of the principal statistical methods used to measure the program net impacts